NWDAF Processing Entity Analytics for Split AI/ML Offloading

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing network performance analytics provided by NWDAF do not consider architecture, available capability, and energy consumption in the processing network, leading to inaccurate assistance for AI/ML operations requiring task offloading between UE and the network.

Innovation Solution

Implement methods for NWDAF to compute and provide PE performance analytics, including statistics and predictions on resource availability, communication performance, and energy consumption, to assist in splitting AI/ML operations between UE and processing network.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Use of energy by moving object

If AI/ML operations are offloaded to the processing network, then computation resources and energy consumption are improved, but network performance analytics accuracy deteriorates due to lack of consideration for architecture, capability, and energy consumption

Engineering Contradiction:

Improveenergy consumptionVSAvoidnetwork performance analytics accuracy

Core Design Contradiction:

Use of energy by moving objectVSMeasurement precision

Solution Approach 1:

The patent implements feedback mechanisms where the NWDAF continuously collects performance data from the processing network, analyzes it, and uses the insights to improve future task offloading decisions. This closed-loop system ensures that analytics become increasingly accurate by incorporating real-world performance feedback, directly resolving the contradiction between energy-efficient offloading and analytics precision.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system performs preliminary analysis of network architecture, capability, and energy consumption characteristics before making task offloading decisions. By pre-characterizing the processing network's performance attributes and storing this information for reference, the system can make accurate analytics-driven decisions without requiring real-time measurements, thus maintaining analytics accuracy while enabling energy-efficient offloading.

Inventive Principle:

Principle #10Preliminary action

2Productivity

If AI/ML operations are split between UE and processing network, then application performance is improved, but decision accuracy for task offloading deteriorates without accurate performance analytics

Engineering Contradiction:

Improveapplication performanceVSAvoidtask offloading decision accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The NWDAF implements feedback loops that collect actual task execution performance, energy consumption, and latency data from both UE and processing network. This feedback is used to continuously refine the analytics models, ensuring that task offloading decisions become increasingly accurate while maintaining high application performance through optimized split operations.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent replaces ad-hoc, rule-based task offloading decisions with analytics-driven intelligent decision-making. By substituting simple mechanical decision rules with sophisticated analytics models that consider network architecture, capability, and energy consumption, the system achieves both high application performance and accurate task offloading decisions.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Measurement precision

If comprehensive performance analytics are collected from processing network, then task offloading accuracy is improved, but system complexity increases

Engineering Contradiction:

Improvetask offloading accuracyVSAvoidsystem complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the performance analytics collection and processing functions into distinct modular components within the NWDAF. Different analytics modules handle specific aspects such as network architecture analysis, capability assessment, and energy consumption monitoring independently. This segmentation reduces system complexity by organizing comprehensive analytics into manageable, reusable modules while maintaining high task offloading accuracy.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The NWDAF is designed as a universal analytics function that can serve multiple purposes: task offloading decisions, resource allocation optimization, energy management, and performance monitoring. By creating a multi-functional analytics platform, the system achieves comprehensive task offloading accuracy without proportionally increasing complexity, as the same analytics infrastructure supports multiple objectives.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS20260067170A1Performance analytics for assisting machine learning in a communications network

Publication Date: 2026.03.05 TELEFONAKTIEBOLAGET LM ERICSSON (PUBL)

AI summary

A method for a network data analytics function, NWDAF, configured to assist splitting of artificial intelligence/machine learning, AI/ML, operations between a user equipment, UE, and a processing network that are operably coupled via the communication network, the method comprising receiving, from a network function, NF, or an application function, AF, associated with the communication network, a request for a processing entity, PE, performance analytic associated with the processing network, wherein the processing network comprises a plurality of PEs; for each PE in the processing network, obtaining one or more of the following information: PE resource availability, and communication performance between the PE and each other PE in the processing network; computing the PE performance analytic based on the obtained information; and sending the computed PE performance analytic to the NF or AF, in accordance with the request.